Machine Learning in Hardware Security of IoT Nodes
T Lavanya, K. Rajalakshmi · 2021
Machine learning algorithms are in wide use, and they have been utilized by attackers and defenders. Attackers use them to attack by sending an unwanted signal to a circuit, and defenders detect the presence of an unwanted signal using machine learning. Thus, we are going to identify the effect of machine learning on hardware security, i.e. we are going to realize the machine-learning–based mechanism of attack and defense for hardware. As we know, there is a widespread Internet of Things (IoTs), and this is a growing network of devices that establishes Internet connectivity and communication among themselves. The IoT uses billions of data as it connects billions of devices to the Internet. Due to this expansion, IoTs are vulnerable to attacks by adversaries. Adversary attacks cause malicious functions like changes in the functionality of circuits, leakage of data, reduced reliability, modification of physical parameters, etc. Hence, the security of IoT devices has become a major concern. Thus, for each hardware security issue we need to identify the suitable machine learning algorithm. We are going to highlight the important aspects of machine learning applications over hardware security problems.